The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions for installing the gtars Python package and gtars-cli command-line tool using standard package managers (uv pip and cargo). These are routine operations for installing legitimate software from public registries.
[COMMAND_EXECUTION]: The skill documentation includes numerous examples of using the gtars CLI for specialized genomic tasks, including building IGD indexes, generating coverage tracks, and splitting fragment files. The commands are consistent with the skill's stated purpose and do not perform unexpected operations.
[PROMPT_INJECTION]: The skill facilitates the ingestion and processing of external genomic data files (BED, FASTA, TSV), which creates a potential surface for indirect prompt injection. 1. Ingestion points: Untrusted data enters the agent context through methods such as RegionSet.from_bed (SKILL.md, python-api.md), RefgetStore.from_fasta (SKILL.md, refget.md), and CLI inputs (cli.md). 2. Boundary markers: The instructions do not define specific delimiters or guardrails for the content of these genomic files. 3. Capability inventory: The skill possesses file-writing capabilities (to_bed, to_json) and performs network retrieval via the bbcache fetch command (cli.md). 4. Sanitization: There is no mention of sanitizing or validating the contents of ingested genomic data beyond standard format parsing.

gtars